Face Detection and Tracking Using Raspberry Pi based on Haar Cascade Classifier

Abstract

Face detection and tracking technology is the key link for intelligent service robot to complete visual perception and human-robot interaction. This paper research the problems of slow speed and low accuracy for face detection and tracking on the low performance of embedded devices. In this paper, face detection and tracking are implemented on a smart car controlled by Raspberry Pi 4B. The face detection algorithm used in this research is the haar cascade classifier method in the OpenCV library which is run using the python programming languages, and the scaling process is used to the image to speed up the detection speed and improve the frame rate. Face tracking is based on the idea of keeping the detected face frame in the central area of the detection window, the coordinates of the center of the face frame in the X direction are calculated firstly, and the corresponding tracking commands are executed by judging the range of the coordinates. Experiments show that the face detection and tracking of the Raspberry Pi car based on the haar cascade classifier method in this paper can identify and locate faces in real time and accurately, and can effectively track the detected faces.